A Danish study allocated at random 30,970 persons to screening with Hemoccult-II and 30,968 as controls in a population of 140,000 between 45 and 74 years old. Persons with known colorectal cancer, adenoma, and distant spread from all types of cancer were excluded. The test was completed in 20,672 persons from August 1985 to September 1986, and 215 (1%) were found to be positive. Colonoscopy in 203 and double-contrast barium enema in 6 detected 37 persons with cancer and 86 with adenomas. Dukes A cancer was detected in 19 in the screening group, in contrast to 2 among controls. Synchronous adenomas were found in 23 with a positive test and 10 controls. Interval cancer was found in nine persons within 1-11 months after a negative test. Eighteen persons got cancer before invitation and six defectors as well. In all, colorectal cancer was detected in 70 persons in the screening group and in 38 controls; the figures for adenomas alone were 103 and 38, respectively. The study is designed to detect a possible reduction in mortality from colorectal cancer of 25% within 5 years after three screenings with intervals of 2 years. The second screening will begin in September 1987. It remains to be shown whether a reduction may be obtained both in mortality and in incidence of colorectal cancer.